Hugging Face's logo

RESMP-DEV
/
LFM2.5-Encoder-350M-MLX-FP32

MLX
Safetensors
lfm2
encoder
liquidai
fp32
custom_code
Model card Files
xet
 Community
LFM2.5-Encoder-350M-MLX-FP32
File size: 2,365 Bytes
591cda4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
{
  "source": "LiquidAI/LFM2.5-Encoder-350M",
  "source_revision": "b886781f7c6f10ca9b7096e21b83e30a073c2f39",
  "runtime": {
    "mlx_device": "Device(gpu, 0)",
    "torch_device": "cpu",
    "dtype": "float32"
  },
  "tensor_audit": {
    "tensor_count": 148,
    "exact_tensor_count": 148,
    "parameter_count": 354483968,
    "all_source_tensors_fp32": true,
    "all_mlx_tensors_fp32": true,
    "all_values_bit_exact_after_layout_transform": true,
    "source_weights_sha256": "cd70e404c3c6c1756b2cf5dc75de2a87788b460e3a6156c1a4ab134b824c2706",
    "mlx_weights_sha256": "59318ef252f74691ff6a5338f1009c9723a91126f6e9f4361f0bdfa41b98c886"
  },
  "parity": {
    "input_shape": [
      3,
      13
    ],
    "valid_token_count": 31,
    "hidden": {
      "max_abs_error": 0.0003604888916015625,
      "mean_abs_error": 3.0134134043696157e-06,
      "rmse": 5.094438406081536e-06,
      "relative_l2_error": 4.034416993567579e-06,
      "cosine_similarity": 0.9999999999919315
    },
    "logits": {
      "max_abs_error": 0.0003662109375,
      "mean_abs_error": 1.4175360927475862e-05,
      "rmse": 2.2781126167816667e-05,
      "relative_l2_error": 2.509927860453007e-06,
      "cosine_similarity": 0.9999999999968735
    },
    "top1_agreement_valid_tokens": 1.0,
    "top1_agreement_masked_tokens": 1.0,
    "masked_predictions": [
      {
        "text": "The capital of France is <|mask|>.",
        "reference_token_id": 5242,
        "mlx_token_id": 5242,
        "reference_token": " Paris",
        "mlx_token": " Paris",
        "match": true
      },
      {
        "text": "Liquid neural networks run efficiently on <|mask|> devices.",
        "reference_token_id": 64014,
        "mlx_token_id": 64014,
        "reference_token": "python",
        "mlx_token": "python",
        "match": true
      },
      {
        "text": "Paris is the <|mask|> of France.",
        "reference_token_id": 5706,
        "mlx_token_id": 5706,
        "reference_token": " capital",
        "mlx_token": " capital",
        "match": true
      }
    ]
  },
  "acceptance": {
    "passed": true,
    "checks": {
      "bit_exact_fp32_weights": true,
      "hidden_relative_l2_below_1e-4": true,
      "logits_relative_l2_below_1e-4": true,
      "valid_token_top1_agreement_100_percent": true,
      "masked_token_top1_agreement_100_percent": true
    }
  }
}